Train 1D CNN

1D Convolution을 기본 구성 요소로 하는 EEG classifier를 학습해보는 노트북.


환경 구성


Data Filtering by Diagnosis

Non-Vascular Dementia, Non-Vascular MCI, Normal


Configure the Train, Validation, and Test Splits

Split the filtered dataset and shuffle them

Wrap the splitted data using PyTorch Dataset

Data loader test

Train, validation, test dataloaders


Define 1D CNN models

1D Tiny CNN

M5-like model

M5-like model without the usage of age

1D ResNet model

1D CNN: ResNet-like model without the usage of age

1D Tiny ResNet


Some useful functions for training


Train models